Idea Intelligence · b2b
CompassPoint
Real-time compensation benchmarking platform with AI market intelligence that helps companies pay competitively and retain top talent
The problem
Compensation decisions are being made with stale data in a market that moves in real time. The traditional compensation benchmarking cycle (purchase an annual survey from Radford or Mercer, analyze the data six months after collection, set salary bands for the next fiscal year) is structurally broken in a world where engineering salaries moved 20% in 12 months during 2021-2022 and then compressed 15% in 2023. By the time survey data reaches HR teams, it is 12-18 months old and already obsolete. The consequence is brutal: companies discover they are underpaying top performers only when resignation letters arrive, at which point counter-offering costs 20-30% more than proactive adjustment would have. Pay equity liability accumulates invisibly, companies with gender or ethnic pay gaps face regulatory risk and reputational exposure that grows with each hiring cycle. Compensation bands for new roles (machine learning engineer, AI product manager, prompt engineer) are nonexistent in legacy survey instruments that update annually. Finance teams cannot model total compensation cost accurately without real-time data, leading to budget variances that surprise boards. The market for real-time compensation intelligence exists but no platform delivers it at the speed, granularity, and AI depth that modern HR teams require.
The solution
CompassPoint delivers a real-time compensation intelligence layer built on three data streams: aggregated job posting salary data from 50+ job boards updated daily, anonymized offer letter benchmarks contributed by participating companies, and traditional survey data from Radford, Mercer, and Willis Towers Watson mapped into a unified taxonomy. The AI layer does what spreadsheets cannot: it detects emerging compensation trends by role and market 6-9 months before they appear in annual surveys, flags individual employees whose compensation is below market percentile thresholds and at retention risk, identifies pay equity gaps across gender and ethnicity, and models the cost of bringing the workforce to target percentiles. Compensation analysts receive role-specific market ranges at P25/P50/P75/P90 updated weekly rather than annually. Finance teams get a total compensation cost model that updates with market movements. HR business partners receive an employee-level risk dashboard showing which team members are most likely to receive competitive offers based on market movement in their role and skills cluster.
Why now
Two developments in 2024-2025 make this moment decisive. First, pay transparency legislation has reached critical mass: 13 US states now require salary ranges in job postings as of 2025, up from 2 in 2022. The EU Pay Transparency Directive requires companies to publish pay band data starting in 2026. This legislation does two things simultaneously, it floods the market with salary signal data that powers real-time benchmarking, and it creates compliance risk for companies without defensible compensation frameworks. Every employer posting jobs in California, New York, Colorado, or Washington is already generating the data that CompassPoint can aggregate. Second, the AI talent market reset of 2024 created unprecedented compensation volatility for AI and ML roles: prompt engineers, AI product managers, and ML infrastructure engineers saw salary ranges shift by 25-40% within 18 months, making annual surveys useless for these critical, expensive roles. HR teams that relied on 2022 Radford data for 2024 AI hiring decisions drastically overpaid or underpaid, both outcomes are costly.
The moat
CompassPoint's moat is its proprietary data consortium. Participating companies contribute anonymized offer letter data in exchange for access to the platform, a data network effect where each new participant improves benchmark accuracy for all others. This consortium data has verification and recency qualities that public job posting data and self-reported data cannot match. The AI models trained on this proprietary dataset accumulate predictive accuracy that improves with scale, a classic flywheel that widens the competitive gap over time. Compensation taxonomy investment is also a durable barrier: mapping 10,000+ job titles across industries and geographies into a consistent framework requires years of expert curation and is not replicable quickly. Deep HRIS integrations (Workday, SAP SuccessFactors, Oracle HCM) that pull live employee data for retention risk modeling create switching costs, once CompassPoint is embedded in compensation cycle workflows, migration is painful. Pay equity analysis features tied to regulatory reporting create compliance dependencies that sustain renewal conversations.
How it makes money
Subscription pricing anchored to company employee count and number of markets monitored. Mid-Market tier for 200-1,000 employees at $1,500/month covers up to 5 geographies and 50 job families with weekly benchmark updates and pay equity reporting. Enterprise tier for 1,000-10,000 employees at $4,000/month adds full compensation band modeling, employee retention risk dashboard, HRIS integration, and offer analysis workflows. Global Enterprise at $9,000+/month covers unlimited geographies with custom job taxonomy mapping, dedicated customer success, and regulatory reporting modules for EU pay transparency compliance. Annual contracts with a 20% discount are standard. Professional services (job family architecture design, compensation philosophy development, pay equity remediation planning) provide one-time project revenue at $500-750/hour. Data consortium participation is free for employers who share anonymized offer data, creating an acquisition channel. Target gross margin of 75% at scale.
How you'd build it
Months 1-3 build the data pipeline: scrape and normalize salary range data from job postings across Indeed, LinkedIn, Glassdoor, and Lever at daily frequency. Develop job taxonomy mapping 5,000+ titles into a consistent classification framework. Build basic benchmarking UI: role search, market range display, geographic filter. Recruit 10 design partner companies to contribute offer letter data in exchange for free access. Months 4-6 develop the pay equity analysis module, gap detection, statistical significance testing, remediation cost modeling. Build Workday and BambooHR HRIS integrations for employee data sync. Launch the compensation band management tool. Months 7-9 develop the AI retention risk model using combined compensation gap data and role-level market movement. Build the offer analysis workflow: a tool that evaluates a candidate offer against real-time market data and recommends an approval or adjustment. Launch EU pay transparency reporting module ahead of 2026 directive. Months 10-12 expand HRIS integrations to SAP SuccessFactors and Oracle HCM. Build the data consortium contribution workflow for offer letter data sharing. Target 60 paying customers with $1.2M ARR by end of year one.
Proof signals
Levels.fyi, a crowd-sourced compensation database for tech roles, attracts 5M+ monthly users and generates $30M+ in annual revenue despite offering no employer-facing product, proving massive demand for real-time salary data. Payscale's continued growth and Radford's acquisition into AON demonstrate incumbent market value, while their continued reliance on annual survey cycles signals the opportunity for disruption. Carta launched a compensation benchmarking product in 2024 using its equity data network, validating the real-time data aggregation approach. Pay transparency legislation has generated a tsunami of structured salary data: in 2024, 48% of US job postings now include a salary range, compared to 18% in 2021, creating an unprecedentedly rich public dataset. SHRM's 2025 survey found that 67% of Total Rewards leaders consider their current benchmarking data inadequate for real-time market decisions, and 54% want tools that identify individual employee retention risk.
Cite this. Cancel Atlas Idea Intelligence (2026). “CompassPoint.” https://www.cancelatlas.com/ideas/compasspoint (CC BY-SA 4.0). Concept-stage analysis; projections are illustrative, not financial advice.